--- title: "Visualizing the Aftermath of the Eaton and Palisades Fires" description: "Using satellite imagery to visualize fire scars" author: name: Joshua Ferrer-Lozano affiliation: Master of Environmental Data Science categories: [Python, MEDS, data-viz, Geospatial] date: 12-03-2025 format: html jupyter: eds220-env draft: false ---

You can access the full analysis of the project in this Github repository

About

In January 2025, the Eaton and Palisades Fires broke out in Los Angeles County, burning more than 16,000 structures, forcing mass evacuations, and resulting in at least 28 fatalities (Phillips, 2025). Beyond the immediate destruction, these fires left long‑lasting ecological and social impacts. Understanding both the burn severity and the social vulnerability of affected communities is crucial for disaster response and recovery planning.

This blog post combines satellite imagery analysis and social vulnerability data to examine the fires. Using Landsat 8 surface reflectance data, I created true‑color and false‑color composites to visualize burn scars and vegetation loss. I then integrated the Environmental Justice Index (EJI) dataset to assess how wildfire exposure overlapped with vulnerable populations, focusing on minority populations. Together, these analyses highlight how geospatial data science can reveal both ecological damage and social disparities.

Highlights of the analysis

• Processed fire perimeter shapefiles with GeoPandas to ensure consistent CRS and geometry.

• Restored CRS and handled missing values in Landsat 8 imagery.

• Created true‑color composites for baseline visualization and false‑color composites to highlight burn scars.

• Overlaid fire boundaries on imagery to confirm spatial alignment.

• Clipped EJI census tracts to fire perimeters and mapped elderly population percentages.

Dataset descriptions

Eaton and Palisades Fire Perimeter Data:

The Eaton and Palisades dissolved Fire Perimeter (2025) data published by Los Angeles County, provides shapefiles of dissolved polygons of the Eaton and Palisades fire perimeters. These files of which were daily perimeter snapshots, were merged to create a single fire perimeter for each fire. This dataset was from the NIFC FIRIS fire service and is hosted on ArcGIS Hub containing data for both fires. You can access the data here.

Landsat 8 Data:

Landsat 8 Data NetCDF dataset containing atmospherically corrected reflectance data from Landsat 8 via the Microsoft Planetary Computer. Includes spectral bands (red, green, blue, near-infrared, shortwave infrared) clipped to the Eaton and Palisades fire perimeters. You can access the data here.

Environmental Justice Index (EJI) Data:

Census-tract level socioeconomic and demographic indicators estimating vulnerability, recovery potential, and mitigation capacity. Variables include population, housing, transportation, and socioeconomic status. You can access the data here.

Step 1: Import Libraries & Load Data

Setup

Load the needed packages, import fire perimeters, Landsat bands and EJI data.

Load in Libraries

# Import required libraries for geospatial analysis and visualization
import os
import numpy as np
import rioxarray as rio
import matplotlib.pyplot as plt
import geopandas as gpd
import xarray as xr
# Activate and Import netCDF4
import netCDF4
# Check activation and version
print(netCDF4.__version__)

1.7.3

Load Fire Perimeter Data

Fire perimeters provide the spatial boundaries of interest for the analysis.

# Load fire perimeter shapefiles
eaton = gpd.read_file('data\eaton_fire\Eaton_Perimeter_20250121.shp')

# Load fire perimeter shapefiles
palisades = gpd.read_file('data\palisades_fire\Palisades_Perimeter_20250121.shp')

Load and Landsat Data

The Landsat 8 data contains multiple spectral bands capturing different wavelengths of reflected light. Each band provides information about surface conditions.

# Generate variable for path
NC_fp = os.path.join('data','landsat8-2025-02-23-palisades-eaton.nc')

# Use netCDF4
landsat = xr.open_dataset("data/landsat8-2025-02-23-palisades-eaton.nc", engine="netcdf4")

Load EJI Data

# Read in the EJI dataset
ca_eji_fp = 'data/EJI_2024_California/EJI_2024_California.gdb'
ca_eji = gpd.read_file(ca_eji_fp)

Prepare Landsat Data

Restoring Spatial Information

Restoring Geospatial Information The Landsat dataset includes a spatial_ref variable but does not automatically recognize its coordinate reference system (CRS). We restored the CRS to EPSG:32611 (WGS 84 / UTM zone 11N) to ensure proper alignment with fire perimeter polygons.

Clouded areas contained missing values (NaNs), which can cause visualization errors. To address this, we replaced NaNs with zeros, allowing the imagery to render cleanly without gaps.

# Recover the geospatial information from spatial_ref
landsat = landsat.rio.write_crs(landsat.spatial_ref.crs_wkt)

# Identify which bands have nan values using numpy.isnan()
nan_bands = []
for band in ['red', 'green', 'blue', 'nir08', 'swir22']:
    if np.isnan(landsat[band]).any():
        nan_bands.append(band)

print("Bands with NaN values:", nan_bands)

# Substitute any nan values for 0
landsat = landsat.fillna(0)

Bands with NaN values: ['green', 'blue']

Plotting a True Color Visualization

A true‑color composite combines the red, green, and blue bands to produce a natural‑looking image of the fire area. This visualization provides a baseline view of the landscape as it would appear to the human eye, helping establish context before moving to false‑color analysis.

# Re-plot with the true color image
landsat[['red', 'green', 'blue']].to_array().plot.imshow(robust = True)
plt.title('Landsat 8 True Color Composite (Feb 23, 2025)')
plt.axis('off')
plt.show()

The robust=True parameter ensures that extreme outliers (such as bright clouds) do not distort the color scaling.

False Color Analysis

False Color Composite False‑color imagery uses non‑visible wavelengths to highlight features that are difficult to discern in true color. By assigning SWIR → Red, NIR → Green, and Red → Blue, we create a composite where:

Healthy vegetation appears bright green.

Burned areas and bare soil appear reddish‑brown.

Urban areas appear in neutral tones.

This visualization makes fire impacts much more visible than true color imagery.

# Specify bands for false color composite
landsat[['swir22', 'nir08', 'red']].to_array().plot.imshow(robust = True)
plt.title('False Color Composite (SWIR/NIR/Red)')
plt.show()

Visualize False Color Image with Fire Perimeters

To confirm spatial alignment, we reprojected the fire perimeter shapefiles to match the Landsat CRS. Overlaying these boundaries on the false‑color composite allows us to directly compare burn scars with official fire data.

In the visualization, the Eaton Fire perimeter is outlined in red and the Palisades Fire in blue, clearly delineating the affected areas.

# Reproject fire perimeters to match Landsat CRS
eaton = eaton.to_crs(landsat.rio.crs)
palisades = palisades.to_crs(landsat.rio.crs)

# Create comprehensive map
fig, ax = plt.subplots(figsize=(10, 10))

# Plot false color composite
landsat[['swir22', 'nir08', 'red']].to_array().plot.imshow(
    ax=ax,
    robust=True,
    add_colorbar=False
)
# Add Palisades and Eaton fire perimeters
eaton.boundary.plot(ax=ax, edgecolor='blue', linewidth=2)
palisades.boundary.plot(ax=ax, edgecolor='red', linewidth=2)


# Add false color raster
landsat[["swir22", "nir08", "red"]].to_array().plot.imshow(robust=True)

# Add title
ax.set_title('Landsat 8 False Color Image of the Palisades and Eaton Fire Perimeters')

# Add labels with gray background
plt.figtext(0.20, 0.51, "Palisades Fire", color='red', fontsize=12, bbox=dict(facecolor='white'))
plt.figtext(0.80, 0.50, "Eaton Fire", color='red', fontsize=12, bbox=dict(facecolor='white'))

# Remove axes
ax.axis('off')

# Display the map
plt.tight_layout()
plt.show()

Figure Interpretation

The Eaton fire (overlaid in blue) and the Palisades fire (overlaid in red) outline the burn perimeter. This false color composite displays the Eaton and Palisades fire areas captured by Landsat 8 on February 23, 2025. The image uses shortwave infrared (SWIR), near-infrared (NIR), and red bands to highlight burn scars and vegetation health.

Key features:

• Bright green areas: Healthy vegetation (high NIR reflectance)
• Reddish-brown areas: Burned areas and bare soil (low NIR, moderate SWIR)
• Urban/developed areas: Neutral colors (moderate reflectance across bands)

Social Vulnerability (EJI)

This analysis examines socioeconomic characteristics of communities impacted by the fires. While fires affect all habitats, they do not impact all communities equally. We used the Environmental Justice Index (EJI) to identify socioeconomic vulnerability within fire perimeters.

Match the CRS of the EJI data to Landsat data.

# Reproject all datasets to the same crs
eaton = eaton.to_crs(ca_eji.crs)
palisades = palisades.to_crs(ca_eji.crs)

# Clip tracts to the palisades and eaton fire perimeters
eji_palisades_clip = gpd.clip(ca_eji, palisades)
eji_eaton_clip = gpd.clip(ca_eji, eaton)

# Get the EJI tracts intersecting the palisades fire perimeter
eji_palisades_join = gpd.sjoin(ca_eji, palisades, how='inner', predicate='intersects')

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(20, 10))

# Define common scale for comparison
eji_variable = 'EPL_POV200'
#eji_variable = 'E_AGE65'
vmin = min(eji_palisades_clip[eji_variable].min(),
           eji_eaton_clip[eji_variable].min())
vmax = max(eji_palisades_clip[eji_variable].max(),
           eji_eaton_clip[eji_variable].max())


# Eaton Plot
eji_eaton_clip.plot(
    column=eji_variable,
    vmin=vmin, vmax=vmax,
    cmap='YlOrRd',
    legend=False,
    ax=ax2,
    edgecolor='black',
    linewidth=0.5
)
ax2.set_title('Low-Income Population\nEaton Fire Area',
              fontsize=16)
ax2.axis('off')


# Palisades Plot
eji_palisades_clip.plot(
    column=eji_variable,
    vmin=vmin, vmax=vmax,
    cmap='YlOrRd',
    legend=False,
    ax=ax1,
    edgecolor='black',
    linewidth=0.5
)
ax1.set_title('Low-Income Population\nPalisades Fire Area',
              fontsize=16)
ax1.axis('off')

# Overall title
fig.suptitle('Percentile Rank of Low-Income Population in Fire-Affected Areas',
             fontsize=16)

# Shared colorbar
sm = plt.cm.ScalarMappable(
    cmap='YlOrRd',
    norm=plt.Normalize(vmin=vmin, vmax=vmax)
)
cbar_ax = fig.add_axes([0.25, 0.10, 0.5, 0.03])
cbar = fig.colorbar(sm, cax=cbar_ax, orientation='horizontal')
cbar.set_label('Percentile Rank (0-100)', fontsize=16)

plt.tight_layout()
plt.show()

C:\Users\joshu\AppData\Local\Temp\ipykernel_24776\3834148586.py:56: UserWarning: This figure includes Axes that are not compatible with tight_layout, so results might be incorrect.
  plt.tight_layout()

Figure Interpretation

This analysis reveals the socioeconmic vulnerability within each of the fire perimeters. The maps show percentile rankings of low-income population (households earning less than 200% of the federal poverty level) within census tracts. The higher percentile values (dark red) indicate greater economic vulnerability.

Eaton Fire Area:

The burn area shows higher percentile rankings compared to the Palisades area, with large portions in the 50th-70th percentiles and some tracts reaching the 80th-90th percentiles (Altadena). On average, these darker areas demonstrate that fires disproportionately affected communities that are more socioeconomically vulnerable.

Palisades Fire Area:

The Palisades perimeter and its census tracts show predominantly low to moderate percentile rankings (roughly 10th-40th percentiles). This indicates that there was proportionaly less socioeconomically vulnerable communities affected by the Palisades fire.

Environmental Justice Implications:

These contrasting patterns reveal that while fires affect habitats, communities and individuals, those that are socioeconomically disadvantaged are disproportionaly affected due to less expendible income/assets that would otherwise allow for post-fire transition. This analysis does however provide a glimpse of how communities and local governments can:

• Train for Emergency preparedness
• Spot-targeting communities that are most vulnerable
• Plan for adaptation and mitigation strategies to address natural disasters

Conclusion

Wildfires leave both physical and social scars. False-color composites reveal burn severity, while EJI data highlights disparities in who is most affected. The Eaton Fire disproportionately impacted lower-income neighborhoods in Altadena, while the Palisades Fire struck wealthier coastal areas.

By combining xarray, rioxarray, and geopandas, we linked satellite imagery with census-based vulnerability data. This integrated approach shows not only where fires burned, but also which communities face the greatest challenges in recovery. As climate change intensifies wildfire risk, combining physical and social data is essential for equitable disaster response and planning.

References

Palisades and Eaton Dissolved Fire Perimeters. (2025). Fire perimeter shapefiles [Geospatial dataset]. ArcGIS Hub. https://hub.arcgis.com/maps/ad51845ea5fb4eb483bc2a7c38b2370c [Accessed Nov. 24, 2025]

U.S. Geological Survey. (2025). Landsat 8 Collection 2 Level-2 surface reflectance data [Satellite imagery dataset]. Microsoft Planetary Computer. https://planetarycomputer.microsoft.com/dataset/landsat-c2-l2 [Accessed Nov. 24, 2025]

Bennett, M. M., Chen, J. K., Alvarez León, L. F., & Gleason, C. J. (2022). The politics of pixels: A review and agenda for critical remote sensing. Progress in Human Geography, 46(3), 729–752. https://doi.org/10.1177/03091325221074691 [Accessed Nov. 24, 2025]

Galaz García, C., Cawse-Nicholson, K., Frew, A., & Fontenot, R. (2024). EDS 220: Working with environmental datasets [Course materials]. Master of Environmental Data Science, Bren School of Environmental Science & Management, University of California, Santa Barbara. https://meds-eds-220.github.io/MEDS-eds-220-course/ [Accessed Nov. 24, 2025]